Computing devices are frequently used to render (e.g., display, print) electronic documents (EDs). However, rendering an ED is a non-trivial operation. Even in resource rich environments (e.g., multithread systems, large memory systems, etc.) throughput may be low if the rendering operation makes inefficient use of the available computing power. As EDs become more complex, users are always seeking to increase rendering throughput, especially when computing resources are available.
There have been attempts in the past to increase rasterization throughput. For example, one attempt involves partitioning a page of the ED into multiple bands, and then rasterizing the bands in parallel using different computer or arithmetic processing units/cores/threads. However, none of these attempts take into consideration the possible existence of self-contained graphical structures within the ED. Specifically, these past attempts do not intelligently exploit multiple threads in view of these self-contained graphical structures.